The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements a self-learning pattern where it reads from and appends to a local knowledge base, creating a vulnerability to indirect prompt injection.
Ingestion points: references/learnings.md is read during the initialization step to gather accumulated platform knowledge.
Boundary markers: There are no delimiters or explicit instructions to treat the content of references/learnings.md as untrusted data or to ignore embedded instructions within that file.
Capability inventory: The agent has the capability to read from the file system and append new content to the references/learnings.md file.
Sanitization: The skill does not define any validation or sanitization procedures for the data being appended to or read from the learnings file.

sales-upfluence